Download Plasmonic modes of gold nano-particle arrays on thin gold

arXiv:1209.2625v1 [cond-mat.mes-hall] 12 Sep 2012
Plasmonic modes of gold nano-particle arrays
on thin gold films ∗
A. Hohenau†, J. R. Krenn
Institute of Physics, Karl-Franzens University Graz, Austria
September 13, 2012
Abstract
Regular arrays of metal nanoparticles on metal films have tuneable optical
resonances that can be applied for surface enhanced Raman scattering or
biosensing. With the aim of developing more surface selective geometries we
investigate regular gold nanoparticle arrays on 25nm thick gold films, which
allow to excite asymmetric surface plasmon modes featuring a much better
field confinement compared to the symmetric modes used in conventional
surface plasmon resonance setups. By optical extinction spectroscopy we
identify the plasmonic modes sustained by our structures. Furthermore, the
role of thermal treatment of the metal structures is investigated, revealing
the role of modifications in the crystalline structure of gold on the optical
properties.
1
Introduction
Due to their spatially confined electron system, metal nanoparticles exhibit
collective electronic modes known as localized surface plasmons (LSP) [1].
∗
A reviewed and edited version of this manuscript is published in Phys. Status Solidi
RRL 4(10), 256, 2010
†
e-mail: [email protected]
1
(a)
(b)
transmission
Λ h
p
d
hf
Z
Y
X
200nm
extinction log(T/Tf)
0.08
0.06
Λ=200
Λ=250
Λ=300
Λ=350
Λ=400
Λ=500
(d) 900
res. wavelength (nm)
excitation by halogen lamp
(c)
0.04
0.02
0
500 600 700
wavelength / nm
800
(1,0) mode
800
(1,1) mode
700
600
500
localized mode
200
400
array period Λ (nm)
Figure 1: (a) Sketch of the sample and measurement geometries. The samples are illuminated from the glass-substrate side, the transmitted light is
collected by a 2.5x, 0.075 numerical aperture objective and analyzed by a
spectrometer; (b) Exemplary SEM image; (c) Extinction spectra of particle
arrays with varying array periods. The reference for calculating the extinction is taken on the gold film outside the arrays. The thin line depicts the
extinction spectrum of a sample with random particle distribution (see text);
(d) Resonance wavelengths of the extinction peaks vs. array period: experimental values from (c) (symbols) and calculated values (lines) of different
grating orders.
These modes usually occur in the visible spectral range and, if excited resonantly, give rise to spatially highly confined (at the nm scale) optical field
enhancement, responsible for surface enhanced Raman scattering (SERS) [2]
or surface enhanced fluorescence [3] effects. Optical field enhancement and
related surface enhanced effects can be further intensified by electromagnetic
interaction between adjacent nanoparticles [2]. In case of particle ensembles
on top of a metal surface, their mutual electromagnetic interaction is partly
mediated by surface surface plasmon polaritons (SPP) [4]. This interaction
was shown to increase SERS efficiency [5], however, a definite assignment
of the observed optical extinction peaks to distinct LSP/SPP modes is still
missing.
In this work, we investigate the optical modes of regular arrays of gold
nanoparticles fabricated by standard electron beam lithography [6] directly
2
on top of a 25 nm thick gold film (Fig. 1), with array periods of 200 − 500 nm.
The samples are investigated by scanning electron microscopy (SEM) and
optical extinction spectroscopy. The investigation of similar systems was
reported recently [7, 8], however for different parameter regimes giving rise to
different optical modes than those investigated in the following. A systematic
variation of the array periods allows us to identify the observed extinction
peaks. Furthermore, we assess the potential of such geometries for refractive
index sensing.
2
Extinction spectra and peak assignment
Fig. 1(c) depicts the measured extinction spectra of arrays of square gold
nanoparticles [Fig. 1(a,b)] with side length d ∼ 100 nm, height hp = 30 nm
and array periods Λ = 200 − 500 nm on a hf = 25 nm thick gold film. The
overall size of the arrays is 100 × 100 µm. For all array periods we observe
one extinction peak at ∼ 520 nm and, additionally, a second peak which
shifts to larger wavelength for larger array periods Λ. Additionally to these
peaks, the extinction spectra show a rise towards longer wavelength. One
can assume that the shifting peaks are related to some coupling phenomenon
between the particles, while the period independent peak at ∼ 520 nm can
be assigned to the excitation of a resonance localized to the single particles.
The spectral position of the peaks in the measured spectra [Fig. 1(c)] are
plotted in Fig. 1(d) as a function of the array period (symbols). In addition,
we plot the calculated grating coupling resonances to antisymmetric SPP
modes (‘a-modes’, defined by the symmetry of the tangential magnetic field
with respect to the gold-film plane; the field maximum is at the gold-glass
interface) in different grating orders (lines). Due to too small array periods,
symmetric SPP modes cannot be excited. The grating coupling resonances
were calculated by finding the spectral minima of the denominator of the
Fresnel coefficient [9] of an unstructured gold film between glass and air. For
the calculation the exciting light wave was assumed to be laterally modulated
with a periodicity equal to the array period Λ. We used the actual gold film
thickness (25nm), the optical constants of gold from Ref. [10] and a refractive
index of n = 1.46 for the quartzglass substrate.
As we find excellent agreement between experimental and calculated data
we conclude that grating coupling to a-mode SPPs is indeed observed. The
period independent peak at ∼ 520 nm can be assigned to a combination
3
(a)
(b)
(c)
(d)
extinction log(T/Tf)
100nm
0.08
0.06
0.04
0.02
0
450 500 550 600 650 700 750 800
wavelength / nm
Figure 2: (a) SEM images of a particle after fabrication; (b) after annealing
for 30 min at 175◦ C; (c) after 4h at 200◦ C; (d) Extinction spectra corresponding to (a) (black line), (b) (red line) and (c) (green line).
of a vertically oriented dipole LSP resonance located at the nanoparticles
and scattering to high-energy SPP [11]. To support this interpretation, we
fabricated a sample with the same average particle density as the Λ = 500nm
array but with randomly distributed particles [thin black line in Fig. 1(c)].
Indeed, in this case only the localized peak and no grating related peaks are
observed.
3
Influence of thermal treatment
By comparing the spectra in Fig. 1 with those published by Felidj et al. [5],
one observes that the latter show a broad extinction feature at ∼ 650−700nm
instead of the monotonic rise above the grating coupling peaks as observed in
Fig. 1(c), although the geometry of the samples was similar. This discrepancy
could result from thermal treatment of the gold film. When prepared by
vacuum evaporation on glass at room temperature, the gold thin films are
built from crystallites about 20nm large. In the standard lithography process
for fabricating samples as used here (and in Ref. [5]) the gold film is covered
by an electron resist [6] which has to undergo an annealing procedure (175◦ C
for 8 hours), leading to a recrystallization of the gold [12]. In contrast,
the gold nanoparticles on top of the gold film are added later and hence
do not undergo thermal treatment. The differences in the surface structure
4
of gold film and particles due to the different crystallite size is obvious in
the SEM image in Fig. 2(a). The corresponding spectrum [black line in
Fig. 2(d)] shows indeed the same broad peak at as observed in Ref. [5].
Further experiments with particles of different shapes and sizes (not shown),
revealed no significant change of the spectral position of the grating coupling
resonances or the appearance of the broad ∼ 650 − 700 nm peak.
If we now anneal the whole sample for 30 minutes at 175◦ C, recrystallization within the particles occurs while their shape is maintained [13]
[Fig. 2(b)]. The spectra then show only a monotonic increase of the extinction towards the red spectral range [red line in Fig. 2(d) and spectra in
Fig 1]. By further annealing (4h at 200◦ C), the particles loose their shape
and melt into the crystal grains of the substrate [Fig.(2(c)]. In the extinction
spectrum, this leads to a disappearance of the localized peak at ∼ 520 nm
and an increase of the grating coupling peaks [green line in Fig. 2(d)]. We
assign the observed differences upon thermal treatment to changes in the particle’s dielectric function and shape, in accordance with the finding reported
in Ref. [13].
4
Influence of superstratum: potential for sensing applications
Although featuring a field maximum at the gold-glass interface, the a-mode
SPP is of considerable strength at the gold-superstrate interface for thin gold
films (below ∼ 40 nm). Importantly, the a-mode is bound strongly to the
interface, giving superior surface selectivity as compared to the conventional
surface plasmon resonance (SPR [4]) scheme, making it potentially attractive
for refractive index (RI) sensing.
We illustrate this in Fig. 3 by qualitatively comparing the shift of the
calculated resonance positions of our arrays with that of the SPR angle. The
investigated system consists of a standard glass substrate (n = 1.52), a 10nm
(25 nm) thick Au film, an adlayer of varying thickness (n = 1.4) and water
(n = 1.33) as the superstrate. While the slope of the spectral resonance
shift of our grating coupling resonance of the 10 nm (25 nm) film is about a
factor 4 (8) smaller than for the SPR resonance (thin black line in Fig. 3),
the sensing depth (defined as 63% of the maximum peak shift) is ∼ 100 nm
for the SPR scheme but only ∼ 13 nm (∼ 26 nm) for the grating coupling
5
624
76
0.12
622
74
0.1
0.08
72
0.06
500
620
0
550
600
650
vacc. wavelength / nm
0.1
0.05
0.15
adlayer thickness / µm
o
78
resonance angle /
25nm Au, GC
10nm Au, GC
25nm Au, SPR
626
extinction
resonance wavelength / nm
80
628
700
70
0.2
Figure 3: Calculated comparison of the resonance shift for a-mode grating
coupling (GC) and SPR. The array periods are chosen as Λ = 151 nm (Λ =
285 nm) for the 10 nm (25 nm) film to get a resonance at 620 nm. The thin
black line shows the resonance wavelength of the SPR at a constant angle of
incidence of 71.8◦ . The inset depicts the measured extinction spectra of an
array with Λ = 350nm nm and hf = 25 nm in air (black) and covered with
10 nm thick SiO2 adlayer (red).
scheme, i.e., clearly less susceptible to bulk RI changes.
As a proof of principle, we experimentally measured the resonance shift
of one of our arrays (after annealing for 4h at 200◦ C) when adding a 10 nm
thick SiO2 on the array-air side. For the array with Λ = 350 nm we observe
a resonance shift of 9 nm of the (1,0) grating peak (inset of Fig. 3). The
estimated sensitivity of our current setup is ∼ 1 nm SiO2 in air.
5
Conclusion
We demonstrated that the extinction spectra of regular arrays of gold nanoparticles on 25 nm thick gold films on glass substrate are governed by extinction
peaks related to grating coupling to a-mode SPP with a field maximum at
the glass-gold interface, and by a resonance localized to the single particles
at ∼ 520 nm. By annealing of the array the particles fuse with the crystal
structure of the substrate, the localized resonance disappears and the grating coupling is enhanced. Such a geometry was used to demonstrate the
feasibility of sensing a thin dielectric layer on top of the array by monitoring
the resonance wavelength of the grating coupling peaks. The sensitivity is
less than for a usual SPR sensor, the surface selectivity is however increased
6
by a factor of 4-8 due to the better spatial confinement of the SPP near
fields. As an additional advantage we note that due to the grating coupling
scheme there is no need for the substrate refractive index to be larger than
the superstrate index.
Acknowledgement
We acknowledge support from the Austrian Science Foundation FWF under
grant Nr. P21235-N20.
References
[1] U. Kreibig and M. Vollmer. Optical Properties of Metal Clusters. Number 25 in Springer Series in Materials Science. Springer, 1995.
[2] K. Kneipp, M. Moskovits, and H. Kneipp, editors. Surface-Enhanced
Raman Scattering, volume 103 of Topics in Applied Physics. Springer,
Heidelberg, New York, 2006.
[3] M. Moskovits. Surface-enhanced spectroscopy. Rev. Mod. Phys., 57:783,
1985.
[4] H. Raether. Surface Plasmons on Smooth and Rough Surfaces and on
Gratings, volume 111 of Springer tracts in modern physics. Springer
Verlag, Berlin Heidelberg, 1988.
[5] N. Félidj, J. Aubard, G. Lévi, J.R. Krenn, M. Salerno, G. Schider,
B. Lamprecht, A. Leitner, and F.R. Aussenegg. Enhanced substrate
induced coupling in two-dimensional gold nanoparticle arrays. Phys.
Rev. B, 65:075419, 2002.
[6] M.A. McCord and M.J. Rooks. Handbook of Microlithography, Micromachining and Microfabrication, volume 1. SPIE and the Institute of
Electrical Engineers, Bellingham, Washington, 1997.
[7] A. Hohenau, J. R. Krenn, J. Beermann, S. I. Bozhevolnyi, S. G. Rodrigo,
L. Martin-Moreno, and F. Garcia-Vildal. Spectroscopy and nonlinear
microscopy of Au nanoparticle arrays: Experiment and theory. Phys.
Rev. B, 73:155404, 2006.
7
[8] J. Cesario, R. Quidant, G. Badenes, and S. Enoch. Electromagnetic
coupling between a metal nanoparticle grating and a metallic surface.
Opt. Lett., 30:3404, 2005.
[9] J. J. Burke, G. I. Stegmann, and T. Tamir. Surface-polariton like waves
guided by thin, lossy metal films. Phys. Rev. B, 33(8):5186, April 1986.
[10] P.B. Johnson and R.W. Christy. Optical constants of the noble metals.
Phys. Rev. B, 6(12):4370, dec 1972.
[11] W. R. Holland and D. G. Hall. Frequency shifts of an electric-dipole
resonance near a conducting surface. Phys. Rev. Lett., 52:1041–1044,
1984.
[12] D. E. Aspens, E. Kinsbron, and D.D. Bacon. Optical properties of Au:
Sample effects. Phys. Rev. B., 21(8):3290, 1980.
[13] K.-P. Chen, V. P. Drachev, J. D. Borneman, A. V. Kildishev, and V. M.
Shalaev. Drude relaxation rate in grained gold nanoantennas. Nano
Letters, (10):916, 2009.
8